Download Ecological Character Displacement in Adaptive Radiation

vol. 156, supplement
the american naturalist
october 2000
Ecological Character Displacement in Adaptive Radiation
Dolph Schluter*
Zoology Department, the Centre for Biodiversity Research, and
the Peter Wall Institute of Advanced Studies, University of British
Columbia, Vancouver, British Columbia V6T 1Z4, Canada
abstract: I give an overview of the observational and experimental
evidence for ecological character displacement in adaptive radiation.
Sixty-one published cases of character displacement involving closely
related species (congeners) make up the observational data set. All
cases involve divergence, even though parallel and convergent displacement are theoretically possible. Character ratios in sympatry
were greatest when displacement was symmetric (mean 1.54) and
least when asymmetric (mean 1.29), perhaps because the most symmetric resource distributions are also the broadest. Carnivores are
vastly overrepresented in the data compared with other trophic
groups, with herbivores the next most common category. I consider
five hypotheses to explain this pattern, including the possibility that
the likelihood of divergence via competition depends on position in
food webs. Overall, the quality and completeness of observational
data has improved in recent years, as judged by the extent to which
individual cases satisfy six standard criteria. All but one of the criteria
are met in over half the cases. Most often lacking is independent
evidence that the species involved compete for resources. For this
reason, we cannot be sure that divergence in sympatry is usually the
result of resource competition rather than some other interaction.
Field experiments on character displacement, which explore how
interaction strength changes per unit change in phenotypic traits, are
only just beginning. I summarize research on threespine sticklebacks
that used experiments in ponds to test three predictions: that presentday differences between sympatric species are a “ghost” of competition past; that adding a competitor alters natural selection pressures
on a species already present, favoring divergence; and that divergent
natural selection stemming from resource competition is frequency
dependent. In total, the evidence suggests that character displacement
occurs frequently in nature, and it probably plays an important role
in the evolution of diversity in many adaptive radiations.
Keywords: adaptive radiation, character displacement, coevolution of
competitors, divergence, interspecific competition.
Interspecific resource competition has long been regarded
as a major cause of phenotypic differentiation in adaptive
* E-mail: [email protected].
Am. Nat. 2000. Vol. 156, pp. S4–S16. q 2000 by The University of Chicago.
0003-0147/2000/15604S-0002$03.00. All rights reserved.
radiation. The idea is simple: depletion of shared resources
by closely related, morphologically similar species would
favor phenotypes exploiting new resources as yet untapped, leading to divergence. Darwin (1859) referred to
the process many times, but Lack’s (1947) study of the
Galápagos finches really solidified the view. Lack presented
several examples in which beak differences between species
were greater where they occurred together (sympatry) than
where they occurred separately (allopatry). His most memorable case was of the two ground finches Geospiza fuliginosa and Geospiza fortis. Both have similar beak sizes on
islands where each occurs alone, but G. fuliginosa has a
smaller beak and G. fortis a larger beak on islands where
the two coexist. Brown and Wilson (1956) later compiled
a list of similar examples from other birds and dubbed
the phenomenon “character displacement.” These studies
had a profound influence on evolutionary theory, and the
idea that competition greatly influenced the evolution of
diversity soon gained widespread acceptance.
Later views of competition’s role in divergence were
more skeptical. Doubts about its importance spread following a series of critiques that questioned the quality and
completeness of the evidence and the conclusions that
could be drawn from it. For example, Grant (1975) pointed
out that two species with partly overlapping geographic
ranges might exhibit greater differentiation in sympatry
than allopatry if each was responding independently to
the same environmental gradient. Arthur (1982) observed
that, in most of the putative cases of character displacement available at the time, evidence was lacking that differences between sympatric and allopatric character states
had a genetic basis. The patterns might instead have been
the outcome of phenotypic plasticity, which, however interesting, is not character displacement. Developing statistical approaches suggested that morphological differences between species in sympatry were often not greater
than would be expected purely by chance (Strong et al.
1979; Simberloff and Boecklen 1981). Randomly generated
“null” assemblages of species often exhibited patterns similar to those seen in real communities. In other words,
alleged instances of character displacement usually also
had simple alternative explanations, and at that time, few
data sets were complete enough to rule them out.
Character Displacement and Adaptive Radiation S5
The appeal of character displacement remained high,
nevertheless, despite uncertainty over its importance.
While the debates bred skepticism, they also spurred interest in locating fresh examples and in applying rigorous
tests to cases both new and old. The outcome is that 170
cases of ecological character displacement have now been
described. Tests against alternative hypotheses have increased dramatically, and experimental study of the process
has begun at last. Recent overviews of selected subsets of
cases (Taper and Case 1992; Robinson and Wilson 1994)
are much more optimistic about its role in evolution.
Here, I summarize most of the available evidence for
character displacement and its implications. This summary
is based on a literature compilation whose details are given
elsewhere (Schluter 2000). I evaluate the completeness of
the observational evidence and examine it for patterns that
might give clues to the circumstances that favor character
displacement, noting especially its relationship to trophic
level of putative competitors and the magnitude and symmetry of displacement. I also describe research on threespine
sticklebacks in my own laboratory that has attempted to
test experimentally predictions of the character displacement hypothesis. I conclude with a perspective of the importance of character displacement in adaptive radiation.
Definition
I define “ecological character displacement” as the process
of phenotypic evolution in a species generated or maintained by resource competition with one or more coexisting
species. By “resource competition,” I mean the negative
impact of one phenotype on another arising from depletion
of shared resources. Perhaps this definition is too narrow,
as similar character shifts may result from apparent competition (“competition for enemy-free space”), a mutually
detrimental interaction between prey species sharing a predator (Holt 1977; Abrams 2000, in this issue). One can go
further and apply the term to shifts caused by other mutually
antagonistic interactions as well, such as intraguild predation, interspecific killing, and behavioral interference. Here,
I use the term in the more restricted, traditional sense.
Nevertheless, an interesting question is whether resource
competition, rather than one of these other interactions, is
responsible for a given pattern suggesting character displacement. The distinction between competition and other
interactions is blurred if interspecific killing or behavioral
interference are adaptations to minimize interspecific resource competition. I focus on divergent character displacement because this is the most relevant to adaptive radiation.
Convergent and parallel character displacement are theoretically possible (Abrams 1986, 1996), but there are no
examples. I limit attention to studies of closely related spe-
cies (i.e., congeners) because these point to the role of competition at a relatively early stage of diversification.
Kinds of Data
The evidence for character displacement in nature is observational, experimental, and predictive. By “observational,” I mean studies of pattern in existing species assemblages. By “experimental,” I mean experimental tests
of predictions of character displacement, not experiments
testing resource competition per se. Experiments on character displacement include measurement of change in natural selection and phenotypic evolution with change in
the density of competing species or change in the strength
of competition as the phenotype of competitors is varied.
This approach has been little attempted but is, in my view,
the most important future direction of research. The “predictive” approach uses optimality or evolutionary models
to generate quantitative expectations of mean phenotypes
under character displacement and compares them with
expectations from other models that do not include competition. The approach has been used only twice, in Cnemidophorus lizards (Case 1979) and Galápagos ground
finches (Schluter and Grant 1984; Schluter et al. 1985),
and in both cases, the results conformed to predictions of
character displacement. Here, I focus on the observational
evidence and recent experimental results.
Observational Data
The most common pattern detected is exaggerated divergence in sympatry, whereby phenotypic differences between two or more species are greater where the species
coexist (sympatry) than where they occur separately (allopatry). Examples include beak sizes of the two Galápagos
ground finches mentioned earlier, body sizes of solitary
and paired Anolis lizards on small islands of the Lesser
Antilles (Schoener 1970; Losos 1990), and growth performance of sympatric and allopatric stork’s bill Erodium
(Martin and Harding 1981).
The next most common pattern is overdispersion of
trait means (or constant size ratios), whereby the mean
phenotypes of ecologically similar species tend to be evenly
spaced along a size or other phenotypic trait axis. Examples
include canine diameters in the wild Felis cats of Israel
(Dayan et al. 1990), mandible lengths of Odontochila and
Therates tiger beetles (Pearson 1980), and flowering times
of Acacia trees (Stone et al. 1996). No realistic theory of
character displacement would predict size ratios that are
truly constant (Abrams 1990), but competition is regarded
as a likely explanation when species trait means tend to
be overdispersed, provided that ancillary information such
as resource use points the same way.
S6 The American Naturalist
The third and rarest pattern is “species-for-species matching,” defined as unusually similar guild structures or phenotype distributions between sets of species that have
evolved independently. Matching is not the same as convergence, which only requires that extant communities be
more similar today than was the case between their ancestors. Matching does not imply that species phenotypes are
overdispersed within any community (nevertheless, overdispersion may lead to matching). Rather, it implies that
trait means are assigned to species in communities in a
consistent way, implying in turn that the members of each
community interact, perhaps as resource competitors
(Schluter 1990). Examples include the independent evolution of similar sets of ecomorphs in Anolis of large islands
in the Greater Antilles (Williams 1972; Losos et al. 1998),
the repetitive origin of benthic-limnetic species pairs in
fishes of northern lakes (Schluter and McPhail 1993; Robinson and Wilson 1994; Schluter 1996b), and the matched
distributions of body sizes in finchlike birds of Mediterranean habitats (Schluter 1990).
Criteria for Observational Evidence
Six standard criteria listed below have come to be recognized as important for demonstrating character displacement with observational data. These criteria are from
a compilation by Schluter and McPhail (1992) of the most
significant criteria raised by earlier researchers (e.g., Grant
1972, 1975; Strong et al. 1979; Arthur 1982; see also Taper
and Case 1992). I scored studies in the literature according
to which of these criteria have been met: First, phenotypic
differences between populations and species should be
wholly or partly genetic. Second, chance should be ruled
out as an explanation of the pattern. Third, population
and species differences should represent evolutionary
shifts, not just species sorting. Fourth, shifts in resource
use should match changes in morphology or other phenotypic traits. Fifth, environmental differences between
sites of sympatry and allopatry should be controlled. Sixth,
independent evidence should be gained that similar phenotypes compete for resources.
A demonstration of heritable phenotypic variation
within a population is not sufficient to satisfy the first
criterion, which is about between-population differences.
This first criterion is an issue mainly in cases of exaggerated
divergence in sympatry. I assume the criterion is satisfied
in all cases of trait overdispersion and species-for-species
matching because these patterns are based on mean phenotype distributions of species living in the same environment. Even if phenotypic plasticity adjusts species
means in response to niche use and competition, a genetic
component is necessary to maintain differences between
means of different species.
Ruling out chance (second criterion) requires rejection
of a statistical null hypothesis, and this in turn requires
some form of replication. Testing exaggerated divergence
in sympatry requires comparison of means over multiple
independent sympatric and allopatric populations. Trait
overdispersion is tested by comparing the size spacing of
species in a community with that in a “null” distribution
of phenotypes generated by random sampling from a
formal probability distribution (e.g., uniform or lognormal) or from an empirical distribution obtained by randomizing observed trait values. Gotelli and Graves (1996)
review the alternative approaches that have been used
and their underlying assumptions. Species-for-species
matching is tested by asking whether differences between
the phenotype distributions of two or more communities
are significantly smaller than expected in random or randomized assemblages. Replication in the latter two kinds
of tests is provided by the multiple species making up
each community. I ignored the problem that statistical
tests of character displacement are uncorrected for the
nonindependence of replicate populations or species arising from phylogeny or gene flow. The method of Hansen
et al. (2000, in this issue) provides a solution to this
problem for statistical tests of exaggerated divergence in
sympatry.
Character displacement involves evolutionary shifts in
character states in response to competition (third criterion), but biased extinction of phenotypically similar species may yield a similar pattern. For example, if a series
of species pairs is assembled on islands of an archipelago,
and if one of the two species goes extinct on those islands
where the colonists are overly similar, then a pattern is
created in which sympatric species are more different from
each other than are allopatric species. I ruled out biased
extinction if, for example, trait differences in sympatry
exceeded the span among allopatric populations or if mean
phenotypes of member species varied over several communities exhibiting trait overdispersion.
Character displacement arises from depletion of shared
resources (competition), and hence, a connection to resource use must be shown (fourth criterion). Tests of exaggerated divergence in sympatry require a demonstration
of a shift in resource use between sympatry and allopatry.
Tests of species overdispersion and matching require evidence that spacing of resource utilization correlates with
the distribution of mean phenotypes.
Controlling for differences between environment (fifth
criterion) ensures that a pattern suggesting character displacement is not merely the outcome of species evolving
independently to environmental changes (e.g., Grant
1975). It is impossible to rule out all conceivable environmental agents, but attempts should be made to test
effects of the most obvious factors such as resource dis-
Character Displacement and Adaptive Radiation S7
tribution. For example, the evidence for character displacement in Galápagos ground finches is not diminished
when differences between islands in seeds, the finches’
main resource, are controlled (Schluter and Grant 1984;
Schluter et al. 1985). I scored cases of character displacement according to whether or not at least one likely influential environmental factor was controlled. Clearly, this
is not the same as ruling out all environmental differences,
but it is a start.
Finally, a case for character displacement is greatly advanced if the interaction proposed to drive it, resource
competition, is demonstrated (sixth criterion). Such a
demonstration would not rule out the possibility that other
interactions are also present and influencing divergence,
but it would nevertheless satisfy an important precondition for character displacement. I allowed both experimental and observational evidence for competition. Tests
of this criterion represent one area in which experiments
have contributed to the study of otherwise purely observational data. However, such evidence is available in only
a handful of cases (see “Criteria Met”).
These criteria provide a yardstick against which the
completeness of observational evidence for character displacement may be evaluated. Their value lies in the recognition that character displacement is not the only explanation for exaggerated divergence in sympatry, trait
overdispersion, and species-for-species matching. Fulfilling a criterion just means that the corresponding alternative hypothesis has been tested and rejected at least once.
Meeting five of the six criteria is no guarantee that the
sixth too will be upheld. Moreover, satisfying all six criteria
does not prove character displacement—we are dealing
with observational evidence here. Nevertheless, a case is
increasingly compelling as, one by one, the most likely
alternative hypotheses fail upon testing.
densities (Juliano and Lawton 1990a, 1990b). If so, then
resource competition cannot be the mechanism behind
trait overdispersion, and the case was excluded from my
survey (it would nevertheless be interesting to know what
interaction lay behind this pattern, as it may be more
widespread). Thirty-three of the 61 studies included in the
survey were of exaggerated divergence in sympatry, 25
were of trait overdispersion, and three were of species-forspecies matching (Schluter 2000).
Criteria Met
Fulfilling only one of the criteria is not a major feat, but
fortunately such studies are in the minority. Most satisfy
at least three, and 22 (36%) fulfill four or more. A few
cases meet all six criteria. They include threespine sticklebacks (Gasterosteus; Schluter and McPhail 1992; Pritchard 1998), Galápagos ground finches (Geospiza; Schluter
and Grant 1984; Schluter et al. 1985), heteromyid rodents
from deserts of the southwestern United States (Dipodomys, Perognathus, and Chaetodipus; Brown and Munger
1985; Dayan and Simberloff 1994; Heske et al. 1994), fish
species pairs in postglacial lakes (Schluter and McPhail
1993; Robinson and Wilson 1994), and Anolis lizards of
the Greater and Lesser Antilles (Schoener 1970; Williams
1972; Pacala and Roughgarden 1985; Losos et al. 1998).
Which criteria are most often fulfilled and which are
most often lacking? Chance has been ruled out as an explanation of the pattern in 180% of the cases (51 of 61
cases; fig. 1). At the other extreme, independent evidence
that species compete has been gained in only 14 of 61
cases (23%; fig. 1). This evidence is experimental in 10
cases, but even in these, it is not always clear that the
Patterns in Observational Data
Seventy-two cases of character displacement that fulfill at
least one of the six above criteria were found in the literature. They are listed in full in Schluter (2000). Sixtyone of these cases included congeneric species. The latter
set provides the database for the summaries below. The
remaining 11 cases include no congeneric species, and they
are not considered further here. A “case” refers to a unique
pair of species in tests of exaggerated divergence in sympatry, a unique assemblage of species in tests of trait overdispersion, or a unique set of assemblages in tests of
species-for-species matching. The data set does not include
any case that failed one or more tests of the criteria. For
example, body sizes of dytiscid beetles show a significant
pattern of trait overdispersion, but experimental study
suggested that food limitation was absent at natural beetle
Figure 1: Number of observational studies satisfying each of the six
criteria for observational evidence. Data are from Schluter (2000) and
include 61 studies that include closely related species (congeners).
S8 The American Naturalist
mutually antagonistic interaction responsible for the experimental effect is indeed resource competition, not some
other interaction. For example, behavioral interference
may be an important proximate factor in competition experiments (e.g., Gerbillus). Remaining criteria, including
the criterion that the displacement should have a genetic
basis, are fulfilled in roughly half the cases. This represents
a vast improvement since Arthur’s (1982) review.
Character Ratios
Character ratios in sympatry are variable and not always
large (range 1.03–1.98), with a mean ratio of 1.36 (median
1.30). This range is similar to that seen in random assemblages (Eadie et al. 1987). Therefore, the magnitude of
character ratios provides little information about whether
size differences in sympatry show a pattern consistent with
character displacement. Size ratios of adjacent species in
cases of trait overdispersion are smaller on average (1.27)
than in cases of exaggerated divergence in sympatry (1.41).
However, cases of trait overdispersion also represent larger
assemblages than those involving exaggerated divergence
in sympatry (up to 13 species instead of just two), and
this alone is expected to yield smaller ratios, all else being
equal.
Symmetry of Displacement
The symmetry of displacement is quantifiable when measurements from sympatry and allopatry are available for
both members of a pair of species showing exaggerated
divergence in sympatry. According to some theories, asymmetry should be associated with greater divergence and
symmetry with less divergence, all else being equal. This
expectation arises from quantitative genetic theories of
character displacement that found that skew in resource
abundance distributions (i.e., those that peak toward one
extreme endpoint of the resource gradient) tends to facilitate character displacement whereas symmetric resource
distributions tend to hinder it (Slatkin 1980; Milligan
1985). Skewed resource distributions also lead to a coevolutionary equilibrium in which one species exploits the
most frequent resources in both sympatry and allopatry
whereas the other species (the rarer) exploits the peak
resources in allopatry but shifts markedly away from the
peak in sympatry. This theoretical prediction is not necessarily robust because only a few scenarios leading to
asymmetric character displacement have been modeled
and because other parameters besides resource skew may
also influence the magnitude of displacement (e.g., width
of the resource distribution).
Symmetry was computed as the ratio of the displacements for each species, the smaller divided by the larger.
The displacement for each species was calculated as the
natural log of the ratio of its trait mean in sympatry and
its mean in allopatry, the larger divided by the smaller.
Symmetry ranged from 0 (only one of two species shifted
from allopatry to sympatry) to 1 (both species shifted equal
amounts). Data were sufficient to allow calculation of symmetry in only 15 cases. Nevertheless, a pattern emerged
that was opposite in direction to that initially expected
(fig. 2). Character ratios in sympatry were greatest when
displacement was symmetric (mean 1.54) and least for the
asymmetric displacements (mean 1.29; Mann-Whitney
U p 51, P p .006).
A possible explanation for this finding is that resource
breadth, not resource symmetry, is the main determinant
of the magnitude of displacement. Wide resource distributions favor large displacements (Slatkin 1980; Doebeli
1996). To account for the results, the widest resource distributions in nature must then also be more symmetric
than narrower resource distributions. Too little information is available on resources to permit a test of this hypothesis. However, it is telling that several of the largest
and most symmetric displacements are from depauperate
environments: Gasterosteus sticklebacks in northern lakes,
Anolis lizards on islands of the Lesser Antilles, Cnemidophorus lizards on islands in the Sea of Cortez, and Geospiza finches from the Galápagos Islands. The resource gradient may be effectively widest in these circumstances
because few other taxa vie for it.
Figure 2: Character ratios in sympatry in the least symmetric displacements (!0.5) compared with those in the most symmetric displacements
(symmetry 1 0.5). See text for details on the calculation of symmetry.
Data are from Schluter (2000) and are based on 15 studies that include
closely related species (congeners).
Character Displacement and Adaptive Radiation S9
Trophic-Level Effects
The most striking pattern is the vast overrepresentation
of carnivores compared with other trophic groups (fig. 3).
Herbivores (mainly granivores) are the next most frequent
trophic category followed by primary producers (plants).
Cases of character displacement in microbivores and detritivores are scarce. The reasons for an association with
trophic level are unknown. At least five hypotheses come
to mind.
Hypothesis 1. Character displacement is most frequent
at those trophic levels experiencing the strongest and most
frequent interspecific competition. Carnivores experience
stronger competition than other tropic levels because
theirs is the only level limited mainly by resources.
Hypothesis 2. Foods consumed by carnivores are nutritionally substitutable whereas those at lower trophic levels
are not. When resources differ in essential nutrients, as
might be true for herbivores and plants, character displacement is more complex and may include convergence
(Abrams 1987, 1996).
Hypothesis 3. Greater species richness at lower trophic
levels dilutes the outcome of competition between pairs of
species, rendering displacements smaller and less detectable.
Hypothesis 4. Character shifts in sympatry are not mainly
the result of resource competition at all, but rather are
evolutionary responses to intraguild predation, interspecific killing, and other direct antagonisms. Such behaviors
are perhaps most prevalent among carnivores.
Hypothesis 5. Variation between trophic levels in the
frequency of cases may result from simple detection bias.
Character displacement is most easily detected when species differ in easily measured morphological traits that map
linearly onto resource utilization. Perhaps this is more
often true of carnivores than other trophic levels.
Detection bias is probably part of the story. Simple relationships between morphology and resource use in carnivores seem to be overrepresented in the literature compared with species at other trophic levels. Simple linear
relationships are also commonly described in granivores,
which make up the majority of the herbivore cases.
Straightforward relationships between phenotype and resource use are far less frequently described for folivores,
plants, and detritivores. Little evidence is available to test
the other hypotheses. Note, however, that many of the
carnivores in the data are not top predators but are intermediate-level predators that are themselves preyed upon
(by diseases but also by other carnivores), detracting from
the first explanation.
Experimental Tests
The above sections reveal that observational evidence for
character displacement is burgeoning and that the mag-
Figure 3: Frequency of cases of character displacement in species at
different trophic levels in community food webs. “Carnivores” refers to
animals that eat other animals, whether vertebrate or invertebrate. “Herbivores” includes consumers of any plant parts such as seeds, leaves, or
fruits. Data are from Schluter (2000) and are only from studies including
closely related species (congeners).
nitude and frequency of displacement varies greatly with
trophic level and other circumstances. However, the causes
of these patterns are uncertain. In particular, we require
more direct evidence that resource competition is the main
cause of exaggerated divergence in sympatry, trait overdispersion, and species-for-species matching. Information
on mechanism remains the category of evidence in shortest
supply (fig. 1). Experiments have confirmed the existence
of a competitive interaction between species in 10 of 61
cases of character displacement. Yet, field experiments on
competition are not rare (Connell 1983; Schoener 1983;
Gurevitch et al. 1992), a fact that should inspire efforts to
assess resource competition’s role in other putative cases
of character displacement.
There is a further reason to conduct experiments that
goes beyond the desire just to fulfill one more criterion
for character displacement. Experiments allow us to test
stronger predictions of the hypothesis than simply that
species compete. Most important, they allow us to explore how interaction strength has changed during the
putative character displacement sequence and to investigate how natural selection pressures on a species are
altered when a second, closely related species is present
in the environment.
Here I summarize experiments on threespine sticklebacks (Gasterosteus aculeatus species complex) designed to
test three related predictions of the character displacement
hypothesis. Where relevant, I point to comparable results
from other systems. I limit attention to experiments that
S10 The American Naturalist
address observed cases of character displacement in nature.
A different class of experiments uses model laboratory
systems to understand how interaction leads to character
divergence and polymorphism (Helling et al. 1987; Taper
and Case 1992; Rainey and Travisano 1998; Travisano and
Rainey 2000, in this issue).
Sympatric Threespine Sticklebacks
Several experiments were carried out to test the hypothesis
that character displacement promoted morphological and
ecological divergence of sympatric sticklebacks inhabiting
small post-Pleistocene lakes of coastal British Columbia,
Canada. The system has been described in detail in Schluter and McPhail (1992), McPhail (1994), and Schluter
(1996a). Briefly, pairs of species are found in several lakes.
In each case one of the species is limnetic, foraging primarily on zooplankton in the water column, and the other
is benthic, foraging on invertebrates in the sediments or
clinging to vegetation. Each limnetic species is small and
slender and has many long gill rakers and a narrow gape.
Each benthic species is larger and more robust and has
fewer, shorter gill rakers and a wider gape. Solitary stickleback species occurring in nearby lakes similar in size to
those containing pairs are intermediate in morphology
(but closer on average to the benthic than to the limnetic)
and consume both zooplankton and benthic invertebrates.
Morphological differences between populations and species persist in the laboratory, and at least some of these
appear to have a polygenic basis (Hatfield 1997).
The experiments were designed around a historical scenario proposed by McPhail for the origin of species pairs
(Schluter and McPhail 1992; McPhail 1993, 1994). He suggested that each pair formed by a process of double invasion and character displacement (fig. 4A). A key part of
this scenario is that the intermediate form resulting after
the first invasion by the marine ancestor was displaced
toward a more benthic lifesyle when the marines invaded
lakes a second time. The second invader did not evolve
an intermediate phenotype like that attained the first time
but instead retained a shape closer (although not identical)
to that of the marine ancestor. The broad features of this
historical sequence are supported by nuclear genes (allozymes and microsatellites; McPhail 1984, 1992; Taylor and
McPhail 2000) and measurements of salinity tolerance
(Kassen et al. 1995). An alternate scenario involving character displacement and sympatric speciation following a
single colonization is supported by mtDNA results. However, we suspect these results are contaminated by mitochondial gene flow between populations after the second
invasion (Taylor et al. 1997; Taylor and McPhail 2000).
Modifications of this double-invasion scenario that are
also consistent with the molecular data are possible (Kas-
Figure 4: A, Hypothesized double invasion and character displacement
sequence for the origin of stickleback species pairs. Curving arrows at
the right indicate invasion events by the marine ancestor. Symbols indicate the mean position of a population along a resource-based morphological continuum: marine, solid star; intermediate, solid triangle; benthic, solid circle; and limnetic, open circles. B, Experimental treatments
simulate stages in (A) before and after character displacement. Modified
from Pritchard and Schluter (in press).
sen et al. 1995), but they do not critically affect the design
of our experiments.
The experiments were carried out in a series of ponds
on the campus of the University of British Columbia
(Schluter 1994). Each pond is 23 m # 23 m, and its depth
slopes gradually to 3 m in the center. The ponds were
constructed in 1991 and seeded with plants and invertebrates from Paxton Lake on Texada Island, British Columbia (an 11-ha lake used as the source of experimental
limnetic and benthic populations). Ponds are sand lined
and edged with limestone extracted from surface mines
near Paxton Lake. The ponds are not intended to be identical to wild lakes but only to mimic natural conditions
well enough that they may serve as tools to study species
interactions. All invertebrates found in the diets of experimental fish were characteristic of the species in the
wild. Fish predators of sticklebacks are absent in the ponds,
but insect predators are common. Predation by piscivorous
birds (herons, kingfishers, and mergansers) is present but
sporadic. Each pond can sustain thousands of sticklebacks
over multiple generations, and life cycles in the ponds are
similar to those in native lakes (D. Schluter, unpublished
observations).
Experiment 1: The Ghost of Competition Past
This experiment tested the prediction that the intensity of
competition experienced by species should decline as divergence proceeded, yielding descendants whose presentday interactions are a “ghost” of their former strength (see
Character Displacement and Adaptive Radiation S11
Connell 1980). This expectation is a standard assumption
of character displacement theory, but it has rarely been
tested. Pritchard and Schluter (in press) tested the prediction by recreating historical and contemporary stages
of the postulated character displacement sequence, taking
advantage of the availability of the ancestral marine species
in the ocean. The experiment ran in the summer of 1995
and lasted 7 wk. The marine species was the target. It was
placed together with intermediate sticklebacks from
Cranby Lake, Texada Island (pre-displacement treatment),
or benthic sticklebacks from Paxton Lake (post-displacement treatment) in three divided ponds (fig. 4B). We used
the intermediate form from Cranby Lake because the lake
is similar to Paxton Lake in location, size, productivity,
vegetation characteristics, and prey types available. The
Cranby lake population has a low salinity tolerance, like
the present-day benthics, suggesting that it formed around
the time of the first invasion to lakes presently containing
two species (Kassen et al. 1995). The experimental design
held the phenotype of the zooplanktivore constant between
treatments so that any treatment effects may be attributable to changes in the first invader, the form that departs
most from the ancestral state (Schluter and McPhail 1992;
Pritchard and Schluter, in press).
The “ghost” prediction was confirmed. Growth rate and
degree of habitat specialization of the marine species was
higher in the postdisplacement treatment than in the predisplacement treatment, suggesting a decline in the
strength of resource competition through time (fig. 5). An
alternate interpretation is that benthics are indirect mutualists that somehow facilitate growth of marines in the
postdisplacement treatment, perhaps by depleting the density of invertebrate predators and competitors of zooplankton, the main food of marine sticklebacks. Such indirect
effects are rarely considered and may be influential in promoting divergence in many systems. However, this alternative does not accord well with other stickleback experiments in these ponds, which indicate that growth
depression is the inevitable net result of adding sticklebacks
to ponds regardless of phenotype. Our results, therefore,
suggest that competition strength fell as divergence
proceeded.
Only two other studies have compared competition intensity between species in different phenotype combinations (not necessarily representing extremes of a character
displacement sequence). Competition was stronger between the two sympatric species of Anolis lizard on the
Caribbean island of St. Maarten, where morphological and
ecological separation is small, than between two species
on St. Eustatius, where differences in body size and perch
use are greater (Pacala and Roughgarden 1985). Competition between the two salamander species Plethodon jordani and Plethodon glutinosus was greater in the Great
Figure 5: Mean growth of marine sticklebacks in two pond treatments
representing early and late stages of a hypothesized character displacement sequence (fig. 4B). Modified from Pritchard and Schluter (in press).
Smoky Mountains, where altitudinal ranges of the two
species abut without overlapping, than in the Balsam
Mountains, where altitudinal ranges overlap extensively
(Hairston 1980).
Experiment 2: Natural Selection Promotes Divergence
The goal of the second experiment was to test whether
natural selection favored divergence between sympatric
stickleback species (Schluter 1994). This experiment has
been reviewed before (Schluter 1996a) and will not be
elaborated here. I give a brief overview of crucial elements
because the next experiment tests a related expectation
and builds upon the same design.
The experiment tested another prediction stemming
from the double invasion scenario (fig. 4A): that addition
of a zooplanktivore (second invader) should alter natural
selection pressures in a solitary population having an intermediate phenotype, favoring the more benthic phenotypes within it. The target of the experiment was the
intermediate stickleback from Cranby Lake, Texada Island.
This population was hybridized to limnetics and benthics
before introduction to increase its levels of phenotypic
variation and thereby to achieve a more sensitive measurement of changing natural selection pressures. Hybridization further enabled us to mark different trophic phenotypes within the target form using traits (armor and gill
raker number) not susceptible to developmental plasticity
that might otherwise have arisen from shifts in habitat use
and diet. The intermediate population was placed either
S12 The American Naturalist
alone (preinvasion treatment) or with the limnetic species
from Paxton Lake (postinvasion treatment) in two divided
ponds. The density of the target was held constant between
treatments to ensure that any changes in selection detected
may be attributed solely to the addition of the zooplanktivore. As a consequence, the total density of fish was
higher in the preinvasion treatment. The hypothesis of
character displacement makes the prediction that phenotypes in the intermediate population that are most similar to the limnetic species in morphology and habitat use
should be most affected by the addition. The goal of the
experiment was to test this prediction (Schluter 1994). The
experiment was carried out over 3 mo in the summer of
1993.
As predicted by character displacement, addition of the
zooplanktivore differentially depressed the growth rate of
those individuals within the target population that were
closest to the added competitor in morphology and diet
(Schluter 1994). The effect of the addition on target phenotypes diminished gradually with increasing morphological distance away from the limnetic, with no growth depression in the most benthic phenotypes within the target
population. The growth differential was steepest in the
pond where density of the competitor was highest by the
end of the experiment.
for zooplankton was a key component of the events that
transpired after the proposed double invasion or whether
instead any added competitor would have produced the
same effect. The experiment was carried out over 3 mo
in the summer of 1996.
The relationship between growth rate and phenotype
differed between treatments, confirming the expectation
that the impact of species addition on the target was
frequency dependent (fig. 6). Adding limnetics instead
of benthics to a pond side reduced the growth rate of
the most limnetic-like phenotypes within the intermediate population and elevated the growth rate of the most
benthic-like phenotypes. Conversely, adding benthics instead of limnetics depressed growth of the most benthiclike phenotypes and raised the growth rate of the most
limnetic-like phenotypes. Figure 6 shows the relationship
when data from the three ponds are combined, but the
result was the same within each replicate (D. Schluter,
unpublished data). It is not possible to determine which
competitor species, limnetic or benthic, exerted stronger
selection because the experiment included no replicates
in which the intermediate species alone was present. Nor
can we assume that the observed relationship between
growth rate and phenotype in the intermediate-only
treatment in 1993 (0 slope; Schluter 1994) would also
have occurred in 1996. Nevertheless, there is an indi-
Experiment 3: Frequency-Dependent Natural Selection
The last experiment tested a third prediction of the character displacement hypothesis: that natural selection generated by interspecific competition should be frequency
dependent (D. Schluter, unpublished data). I tested the
prediction by determining whether the phenotype of the
added competitor influenced the direction of selection
within the intermediate species. The target of the experiment was again the stickleback from Cranby Lake, manipulated as before by hybridization to increase phenotypic
variation. This species was placed on both sides of three
divided ponds. Limnetics from Paxton Lake were added
to one side of each pond, and benthics from the same lake
were added to the other. The density of the target was
again held constant to ensure that differences in selection
detected between treatments may be attributed to the identity of the added competitor. Because competitors were
added to both sides, total fish densities were now also equal
at the start of the experiment.
Unlike the previous experiment, the treatments of the
current design do not replicate stages in the putative history of stickleback species pairs (cf. fig. 4A). There was
no attempt to make this experiment “historically correct”
as in the case of the previous two experiments. Instead,
my goal was to assess whether the identity of the second
invader made a difference, that is, whether competition
Figure 6: Growth rate (mm/90 d) of different phenotypes of the intermediate species in the presence of the benthic species (solid circle) and
the limnetic species (open circle). The X-axis is a composite index representing trophic variation along the limnetic-benthic axis, with more
benthic-like individuals on the left and the more limnetic-like individuals
on the right. This graph combines data from three ponds, all of which
showed the same pattern. From D. Schluter, unpublished data.
Character Displacement and Adaptive Radiation S13
cation that the target species experienced competition,
and possibly therefore selection, in both treatments of
the current design. Regression lines of growth against
phenotype for the two treatments crossed near the middle
of the range of phenotypes present (fig. 6), with mean
growth rate of the target equal in both treatments. This
contrasts with the previous experiment in which the control and treatment regression lines crossed at the left end
of the phenotype distribution, at the most benthic-like
phenotypes. Therefore, the benthic treatment of the current experiment did not duplicate the control treatment
of the previous experiment, suggesting that the benthic
species as well as the limnetic species impacted the target
but in opposite ways.
Discussion
The observational evidence for character displacement was
extracted from a published body of literature that presents
mainly the positive cases but few of the negative ones.
Furthermore, I have not included negative cases in my
survey, and I have not kept a running tally of examples
that once looked promising but then later failed one or
more tests of the criteria (e.g., the Sitta nuthatches; Grant
1975). Published negative cases are few, but this low count
probably does not adequately represent the frequency of
adaptive radiations in nature in which resource competition played little or no role in divergence. For this reason,
I am unable to estimate the proportion of divergence
events in nature in which significant character displacement occurred. Nevertheless, the positive cases summarized here are sufficiently numerous to indicate that character displacement is not a rare or peculiar phenomenon
in nature. Very likely it is common, and it probably plays
an important part in the evolution of diversity in many
adaptive radiations. A view still encountered in the ecological literature, that little or no evidence exists for character displacement, is unsustainable.
The study of character displacement nevertheless has a
long way to go. Key evidence is still lacking in most of
the cases that have been described, particularly about
mechanisms. In all but a few cases, tests are lacking of
whether species indeed interact as resource competitors.
Establishing the mechanism of interaction is especially crucial now that it is clear other interactions such as “apparent
competition” arising from shared predation (a.k.a. “competition for enemy free space”; Holt and Lawton 1994)
may also lead to character displacement (Brown and Vincent 1992; Abrams 2000, in this issue; Doebeli and Dieckmann 2000, in this issue). The contribution of shared predation and other interactions to divergent selection in
adaptive radiation is completely unstudied experimentally,
yet well worth investigating.
Additionally, the literature on which my survey is based
provides little information, observational or experimental,
of the circumstances that promote character displacement
and those that hinder it. Variation between different
trophic levels in the number of cases is the most striking
indication of context-dependent character displacement
and hints that position in food webs may be an important
determinant. Before we can be sure of this, however, the
influence of simple detection bias must be ruled out. A
discovered link between symmetry and magnitude of displacement also indicates external influences on character
displacement that have yet to be clearly identified. As well,
character displacement theory predicts parallel shifts and
even convergence under more complex scenarios of competition between species (Abrams 1986, 1996), but there
are no examples.
Finally, we still know far too little about the dynamics
of competition and selection. This is partly because we
continue to rely too heavily on observational studies of
pattern for our information. The sticklebacks reveal the
feasibility of experimental studies, and I am optimistic that
considerable progress in the future will come from manipulations of species composition and measurements of
resulting changes in selection pressures and the direction
of evolution. Experiments promise stronger tests of character displacement not just because they implement controls and replication but also because they allow tests of
novel predictions that cannot be tested in any other way.
Assessing how interaction strength changes at different
putative stages of divergence is crucial to understanding
its causes. Experiments also provide measurements of the
basic processes involved in character displacement, including natural selection itself.
This study has focused on the role of competition between close relatives in differentiation. The other side of
competition between relatives is “ecological opportunity,”
an abundance of niches in a region little populated by
competing taxa. Simpson (1953) and other naturalists of
his day viewed ecological opportunity as the major regulator of the rate and extent of adaptive radiation. For
diversification to occur, it was felt that the adaptive landscape of similar niches “must be occupied by organisms
for some reason competitively inferior to the entering
group or it must be empty” (Simpson 1953, p. 207). Most
of the conclusions drawn from studies of character displacement between close relatives also apply to tests of the
role of competing taxa in divergence and adaptive radiation. The quality of evidence for character displacement
between species in different genera, families, or higher taxa
is similar to that between more closely related species but
examples are fewer. Evidence that resource competition is
indeed the mechanism promoting divergence between distantly related species is usually absent. The evidence link-
S14 The American Naturalist
ing phenotypic differentiation with ecological opportunity
is almost entirely qualitative, perhaps because ecological
opportunity is difficult to detect a priori and because detailed measurements have not been made. This is particularly true of examples from contemporary taxa. These
patterns are summarized in greater detail elsewhere (Schluter 2000).
Acknowledgments
I thank P. Abrams, M. Bell, and C. Benkman for their comments on the manuscript. My research is supported by
grants from the Natural Sciences and Engineering Research
Council of Canada and from the Peter Wall Institute for
Advanced Studies at the University of British Columbia.
Literature Cited
Abrams, P. A. 1986. Character displacement and niche shift
analysed using consumer-resource models of competition. Journal of Theoretical Biology 29:107–160.
———. 1987. Alternative models of character displacement. I. Displacement when there is competition for
nutritionally essential resources. Evolution 41:651–661.
———. 1990. Ecological vs. evolutionary responses to
competition. Oikos 57:147–151.
———. 1996. Evolution and the consequences of species
introductions and deletions. Ecology 77:1321–1328.
———. 2000. Character shifts of prey species that share
predators. American Naturalist 156(suppl.):S000–S000.
Abramsky, E., M. L. Rosenzweig, B. Pishow, J. S. Brown,
B. Kotler, and W. A. Mitchell. 1990. Habitat selection:
an experimental field test with two gerbil species. Ecology 71:2358–2369.
Arthur, W. 1982. The evolutionary consequences of interspecific competition. Advances in Ecological Research
12:127–87.
Brown, J. H., and J. C. Munger. 1985. Experimental manipulation of a desert rodent community: food addition
and species removal. Ecology 66:1545–1563.
Brown, J. S., and T. L. Vincent. 1992. Organization of
predator-prey communities as an evolutionary game.
Evolution 46:1269–1283.
Brown, W. L., Jr., and E. O. Wilson. 1956. Character displacement. Systematic Zoology 5:49–64.
Case, T. J. 1979. Character displacement and coevolution
in some Cnemidophorus lizards. Fortschritte der Zoologie 25:235–282.
Connell, J. H. 1980. Diversity and the coevolution of competitors, or the ghost of competition past. Oikos 35:
131–138.
———. 1983. On the prevalence and relative importance
of interspecific competition: evidence from field experiments. American Naturalist 122:661–696.
Darwin, C. R. 1859. On the origin of species by means of
natural selection. John Murray, London.
Dayan, T., and D. Simberloff. 1994. Morphological relationships among coexisting heteromyids: an incisive
dental character. American Naturalist 143:462–477.
Dayan, T., D. Simberloff, E. Tchhernov, and Y. Yom-Tov.
1990. Feline canines: community-wide character displacement among the small cats of Israel. American Naturalist 136:39–60.
Doebeli, M. 1996. An explicit genetic model for ecological
character displacement. Ecology 77:510–520.
Doebeli, M., and U. Dieckmann. 2000. Evolutionary
branching and sympatric speciation caused by different
types of ecological interactions. American Naturalist
156(suppl.):S000–S000.
Eadie, J. M., L. Broekhoven, and P. Colgen. 1987. Size
ratios and artifacts: Hutchinson’s rule revisited. American Naturalist 129:1–17.
Gotelli, N. J., and G. R. Graves. 1996. Null models in
ecology. Smithsonian, Washington, D.C.
Grant, P. R. 1972. Convergent and divergent character displacement. Biological Journal of the Linnean Society 4:
39–68.
———. 1975. The classic case of character displacement.
Evolutionary Biology 8:237–237.
Gurevitch, J., L. L. Morrow, A. Wallace, and J. S. Walsh.
1992. A meta-analysis of competition in field experiments. American Naturalist 140:539–572.
Hairston, N. G. 1980. Evolution under interspecific competition: field experiments on terrestrial salamanders.
Ecology 61:817–826.
Hansen, T. F., W. F. Armbruster, and L. Antonsen. 2000.
Comparative analysis of character displacement and
spatial adaptations as illustrated by the evolution of Dalechampia blossoms. American Naturalist 156(suppl.):
S000–S000.
Hatfield, T. 1997. Genetic divergence in adaptive characters
between sympatric species of sticklebacks. American
Naturalist 149:1009–1029.
Helling, R. B., C. Vargas, and J. Adams. 1987. Evolution
of Escherichia coli during growth in a constant environment. Genetics 116:349–358.
Heske, E. J., J. H. Brown, and S. Mistry. 1994. Long-term
experimental study of a Chihuahuan Desert rodent community: 13 years of competition. Ecology 75:438–445.
Holt, R. D. 1977. Predation, apparent competition, and
the structure of prey communities. Theoretical Population Biology 12:197–229.
Holt, R. D., and J. H. Lawton. 1994. The ecological consequences of shared natural enemies. Annual Review of
Ecology and Systematics 25:495–520.
Juliano, S. A., and J. H. Lawton. 1990a. The relationship
between competition and morphology. I. Morphological
Character Displacement and Adaptive Radiation S15
patterns among co-occurring dyticid beetles. Journal of
Animal Ecology 59:403–419.
———. 1990b. The relationship between competition and
morphology. II. Experiments on co-occurring dyticid
beetles. Journal of Animal Ecology 59:831–848.
Kassen, R., D. Schluter, and J. D. McPhail. 1995. Evolutionary history of threespine sticklebacks (Gasterosteus
spp.) in British Columbia: insights from a physiological
clock. Canadian Journal of Zoology 73:2154–2158.
Lack, D. 1947. Darwin’s finches. Cambridge University
Press, Cambridge.
Losos, J. B. 1990. A phylogenetic analysis of character displacement in Caribbean Anolis lizards. Evolution 44:
1189–1203.
Losos, J. B., T. R. Jackman, A. Larson, K. de Queiroz, and
L. Rodrı́guez-Schettino. 1998. Contingency and determinism in replicated adaptive radiations of island lizards. Science (Washington, D.C.) 279:2115–2118.
Martin, M. M., and J. Harding. 1981. Evidence for the
evolution of competition between two species of annual
plants. Evolution 35:975–987.
McPhail, J. D. 1984. Ecology and evolution of sympatric
sticklebacks (Gasterosteus): morphological and genetic
evidence for a species pair in Enos Lake, British Columbia. Canadian Journal of Zoology 62:1402–1408.
———. 1992. Ecology and evolution of sympatric sticklebacks (Gasterosteus): evidence for genetically divergent
populations in Paxton Lake, Texada Island, British Columbia. Canadian Journal of Zoology 70:361–369.
———. 1993. Ecology and evolution of sympatric sticklebacks (Gasterosteus): origin and evolution of the sympatric pairs. Canadian Journal of Zoology 71:515–523.
———. 1994. Speciation and the evolution of reproductive isolation in the sticklebacks (Gasterosteus) of southwestern British Columbia. Pages 399–437 in M. A. Bell
and S. A. Foster, eds. Evolutionary biology of the threespine stickleback. Oxford University Press, Oxford.
Milligan, B. G. 1985. Evolutionary divergence and character displacement in two phenotypically-variable competing species. Evolution 39:1207–1222.
Pacala, S. W., and J. Roughgarden. 1985. Population experiments with the Anolis lizards of St. Maarten and St.
Eustatius. Ecology 66:129–141.
Pearson, D. L. 1980. Patterns of limiting similarity in tropical forest tiger beetles (Coleoptera: Cicindelidae). Biotropica 12:195–204.
Pritchard, J. 1998. Competition and character displacement in sticklebacks. Ph.D. diss. University of British
Columbia, Vancouver.
Pritchard, J. R., and D. Schluter. In press. Declining interspecific competition during character displacement:
summoning the ghost of competition past. Evolutionary
Ecology Research.
Rainey, P. B., and M. Travisano. 1998. Adaptive radiation
in a heterogeneous environment. Nature (London) 394:
69–72.
Robinson, B. W., and D. S. Wilson. 1994. Character release
and displacement in fishes: a neglected literature. American Naturalist 144:596–627.
Schluter, D. 1990. Species-for-species matching. American
Naturalist 136:560–568.
———. 1994. Experimental evidence that competition
promotes divergence in adaptive radiation. Science
(Washington, D.C.) 266:798–801.
———. 1996a. Ecological causes of adaptive radiation.
American Naturalist 148(suppl.):S40–S64.
———. 1996b. Ecological speciation in postglacial fishes.
Philosophical Transactions of the Royal Society of London B, Biological Sciences 351:807–814.
———. 2000. The ecology of adaptive radiation. Oxford
University Press, Oxford.
Schluter, D., and P. R. Grant. 1984. Determinants of morphological patterns in communities of Darwin’s finches.
American Naturalist 123:175–196.
Schluter, D., and J. D. McPhail. 1992. Ecological character
displacement and speciation in sticklebacks. American
Naturalist 140:85–108.
———. 1993. Character displacement and replicate adaptive radiation. Trends in Ecology & Evolution 8:197–200.
Schluter, D., T. D. Price, and P. R. Grant. 1985. Ecological
character displacement in Darwin’s finches. Science
(Washington, D.C.) 227:1056–1059.
Schoener, T. W. 1970. Size patterns in West Indian Anolis
lizards. II. Correlations with the size of particular sympatric species—displacement and convergence. American Naturalist 104:155–174.
———. 1983. Field experiments on interspecific competition. American Naturalist 122:240–285.
Simberloff, D., and W. Boecklen. 1981. Santa Rosalia reconsidered: size ratios and competition. Evolution 35:
1206–1228.
Simpson, G. G. 1953. The major features of evolution.
Columbia University Press, New York.
Slatkin, M. 1980. Ecological character displacement. Ecology 61:163–177.
Stone, G., P. Willmer, and S. Nee. 1996. Daily partitioning
of pollinators in an African Acacia community. Proceedings of the Royal Society of London B, Biological
Sciences 263:1389–1393.
Strong, D. R., Jr., L. A. Szyska, and D. S. Simberloff. 1979.
Tests of community-wide character displacement against
null hypotheses. Evolution 33:897–913.
Taper, M. L., and T. J. Case. 1992. Coevolution among
competitors. Oxford Surveys in Evolutionary Biology 8:
63–109.
Taylor, E. B., and J. D. McPhail. 2000. Historical contin-
S16 The American Naturalist
gency and ecological determinism interact to prime speciation in sticklebacks, Gasterosteus. Proceedings of the
Royal Society of London B, Biological Sciences (in
press).
Taylor, E. B., J. D. McPhail, and D. Schluter. 1997. History
of ecological seelection in sticklebacks: uniting experimental and phylogenetic approaches. Pages 511–534 in
T. J. Givnish and K. J. Sytsma, eds. Molecular evolution
and adaptive radiation. Cambridge University Press,
Cambridge.
Travisano, M., and P. B. Rainey. 2000. Studies of adaptive
radiation using model microbial systems. American
Naturalist 156(suppl.):S000–S000.
Williams, E. E. 1972. The origin of faunas: evolution of
lizard congeners in a complex island fauna: a trial analysis. Evolutionary Biology 6:47–89.
Associate Editor: Mark A. McPeek